Power Semiconductor Devices for Smart Grid Applications
Lead Research Organisation:
University of Cambridge
Department Name: Engineering
Abstract
The primary goal of the project is to investigate the design and optimization of power semiconductor devices for high efficiency power converters. In particular, improved design features of Schottky diodes and power MOSFETs, which are to be implemented using Silicon carbide (4H-SiC) technology, are to be proposed. This is a natural field and topic of research, due to the substantial reductions in overall unit costs, which could be achieved by means of further progress and enhancement of the aforementioned technology. This would also be in line with the market demand for further advances in the level of hybrid integration of the power, control, communications, and transducer blocks of modern control systems.
Existing commercial 4H-SiC MOSFETs have fallen short of fulfilling their complete potential, due to a plethora of technology limitations. A substantial subset of these are still poorly understood, or haven't been investigated in sufficient depth. That's why this research program also aims to develop a reliable physical model of 4H-SiC devices. Based on this theoretical framework, it would be possible to examine the performance of such structures using both CAD simulators and actual empirical studies. In this way, the project would be able to propose an improved design for a 4H-SiC MOSFET.
Existing commercial 4H-SiC MOSFETs have fallen short of fulfilling their complete potential, due to a plethora of technology limitations. A substantial subset of these are still poorly understood, or haven't been investigated in sufficient depth. That's why this research program also aims to develop a reliable physical model of 4H-SiC devices. Based on this theoretical framework, it would be possible to examine the performance of such structures using both CAD simulators and actual empirical studies. In this way, the project would be able to propose an improved design for a 4H-SiC MOSFET.
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| EP/R513180/1 | 01/10/2018 | 30/09/2023 | |||
| 2275328 | Studentship | EP/R513180/1 | 01/10/2019 | 31/03/2023 | Kaloyan Naydenov |
| Description | The primary findings that have been made as a result of work funded through this award are as follows: - It has been shown that the standard computer models used by both academic and industrial research groups to describe one important phenomenon (Coulombic scattering) in Silicon Carbide power semiconductor devices are largely unphysical and suffer from many limitations. A new model, which is able to resolve these issues, has hence been proposed and has subsequently been applied to various studies performed by our group (including 4 (four) publications, which are publicly available). - The performance and the operation of a new class of Silicon Carbide power devices (junctionless power FinFETs) have been elaborately investigated. It has been demonstrated that although, as some research groups have shown or claimed, these devices are superior to existing technology in some aspects of their operation, they also suffer from various other drawbacks. These limitations have been identified and examined extensively (for the first time) in a journal paper produced by our research group. Based on the results of this investigation, a range of design solutions, which can address the majority of these problems, have been proposed. - As part of our research collaboration with ABB Group and Hitachi Energy, our group has extensively studied the dynamic electro-thermal performance of various power semiconductor devices. In particular, it has been shown (for the first time) that power devices with an advanced layout design (for example, the "hexagonal" layout) can provide important advantages compared to more conventional designs (with a "stripe" layout) in particular aspects of their dynamic operation. This has resulted in three different publications, each of which targets a specific mode of operation of the power device (respectively, "surge current", "short circuit", and "inductive switching" operation). - A comprehensive procedure for the optimisation of a key aspect (called "SRH lifetime profile") of the design of certain types of power semiconductor devices (PiN diodes and MOSFETs) has been proposed for the first time in Silicon Carbide technology. - Through our collaboration with MIRISE Technologies, our research group has been the first to perform fundamental studies on a particular solid state physics effect (which we have named "the FinFET effect") in Silicon Carbide power semiconductor devices. In particular, the significance of this effect and its variation with the operating conditions, the quality of the specimen, etc. have been carefully studied (within the context of the project funded by this award) in the course of two conference papers. In addition, various other findings have also been made as part of our collaborations, yet at present they are still confidential. |
| Exploitation Route | Other researchers can apply or extend the outcomes of this funding by means of some of the following ways (where it must be clarified that, of course, this list is not exhaustive): - Firstly, researchers interested in our work may try to study the computer model (of Coulombic scattering) proposed by our research group in order to evaluate its accuracy compared to the standard models used in both academia and industry. For this purpose, researchers could try to create a full computer model of a given power device and calibrate it to (their) experimental data using either the old expression for Coulombic scattering or the model that we have proposed. Through this comparison, researchers could determine if our model is superior to the standard technique. Therefore, if they do verify that this is the case, they can start to use it in their own studies. This could allow researchers to investigate topics that are otherwise unassailable (from a computer simulation point of view) if the standard model (for Coulombic scattering) is used. Indeed, in the case of our research group, the construction of this model is what allowed us to undertake our studies on the Silicon Carbide FinFET and on the FinFET effect (which would not have been possible to do with any degree of rigour had we continued to rely on the old model). - In addition, our fundamental study on the junctionless power FinFET may convince other researchers of the potential benefits of this device architecture. Thus, they may be motivated to fabricate this device and study it further. In addition, others (including those who originally proposed this power device for Gallium Nitride technology) may potentially find alternative methods to address the key limitations of this power device (identified by our research group) which are more effective than the solutions that we have proposed. - The studies that we have conducted on the FinFET effect in Silicon Carbide power devices could encourage other researchers to further examine this phenomenon (with greater depth and rigour than our works) and even attempt to fabricate and characterise a real device. They could also use our experimental and simulation results to assess the accuracy and the limitations of our computer models and potentially improve them. In addition, similarly to the case of the junctionless FinFET above, they could find other ways to address the issues of this device architecture. Finally, if our work manages to convince other researchers of the large potential of this device (the Silicon Carbide FinFET), they may embrace this technology and develop it into a full commercial product (as our collaborators at MIRISE Technologies are currently trying to do). - In the case of our remaining findings, other researchers could use them to aid the design process for their own power semiconductor devices. |
| Sectors | Electronics,Energy |
| Description | In addition to their academic value, the findings from this award have also had an economic and societal impact due to the industrial focus of this project. In particular, through our work on the study and the design of novel power semiconductor devices, our research group has been able to enhance the technology of our industrial partners by proposing specific designs and design concepts that allow the performance of their technical products to be improved. As a result, the findings from this award have directly contributed to increasing the capabilities and the efficiency of commercial electrical power conversion systems (produced by our collaborators). In addition, in our publications we have tried to study and propose solutions to (or computer models for) various real-world technical challenges that manufacturers of power semiconductor devices experience in general. For example, we have demonstrated that the default computer models used by many companies to describe one key process (Coulombic scattering) in their power devices have severe limitations, and hence we have proposed a new method to describe this phenomenon, which resolves many of those issues. As a result, our collaborators (MIRISE Technologies) now use a calibrated version of this computer model, which our research group has developed, in their simulations. |
| First Year Of Impact | 2020 |
| Sector | Electronics,Energy |
| Impact Types | Societal,Economic |
| Title | A piece of software which provides a memory capability to semiconductor device simulators and allows for the investigation of dynamic, time-dependent changes in the properties of the specimen |
| Description | A new piece of software, which could be used for the purpose of dynamic simulations of semiconductor devices, has been produced. In particular, this "tool" is a set of algorithms that are implemented in a C++ class file and are then combined with a commercial finite element simulator (which has been licensed to our group). The code allows the instantaneous properties of the specimen (during a given simulation) to be stored in memory as the simulation continues and then provides a way for them to affect the simulation results that would be observed in case the same test is performed on the specimen again (after the environment has settled back to its previous state); i.e. this new piece of software provides a memory capability to the simulator, which can now take into account the possibility that the specimen's properties could change dynamically as a result of previous experiments. Commercial semiconductor device simulators do not feature this memory-storage capability; instead, they would predict that the results to all tests would be the same, disregarding the possible impact of past experiments on the properties of the sample that is investigated. As a result, previous simulation-based studies in the field of power semiconductor devices have not been able to reproduce the impact of dynamic effects such as material/device degradation on the current/voltage characteristics of the power device. In this context, the tool, which our group has created, allows the researcher to perform simulations that can directly take into account time-dependent effects such as device degradation. An important application of this technique in my research field could be modelling the phenomenon of forward bias degradation that is observed in bipolar Silicon Carbide power semiconductor devices. In addition, variants of the code could be used to compute various expressions (such as a line integral of a given quantity), which are otherwise not returned to the user by the simulator, in real time (as the simulation is on-going). The results of these computations could then be used to create a (user-defined) model of a given parameter/quantity that the simulator can use in place of the default built-in models, which come along with the simulator and have certain limitations. |
| Type Of Material | Improvements to research infrastructure |
| Year Produced | 2021 |
| Provided To Others? | No |
| Impact | This research tool has so far been used in three separate studies, which are publicly available (one is a conference paper published in the proceedings of ISPSD 2021 (DOI: 10.23919/ISPSD50666.2021.9452282); another is a journal paper (DOI: 10.1088/2631-8695/ac12bc); the final one has just been accepted for presentation and publication in the proceedings of ISPSD 2022). In particular, it has been used to create a new model for the Coulombic mobility in Silicon Carbide power MOSFETs (in place of the default model in the simulator). The use of this new model has then allowed our research group to reproduce the experimental results of our collaborators (MIRISE Technologies) and hence conduct the aforementioned studies. In addition, this new piece of software has also been used to conduct simulation-based studies on the dynamic nature of forward bias degradation in Silicon Carbide bipolar devices as part of my research group's collaboration with Hitachi. The results of these investigations, however, are still confidential. |
| Description | Advanced SiC power semiconductor devices |
| Organisation | ABB Group |
| Department | ABB Corporate Research Centre (Switzerland) |
| Country | Switzerland |
| Sector | Private |
| PI Contribution | The primary contribution of my research group to this collaboration has been studying and evaluating the performance of various Silicon Carbide power semiconductor device designs for the benefit of our industrial partners by means of computer simulations. In this respect, the majority of the corresponding simulation results that have been obtained and the subsequent analysis that has been carried out is confidential and belongs to our collaborators. Nonetheless, the other members of my research group and I have been able to get four conference papers to be published as part of this collaboration. Two of these were presented at ISPSD 2021 and ISPSD 2022 (with respective DOIs of 10.23919/ispsd50666.2021.9452208 and 10.1109/ISPSD49238.2022.9813631), while another two were published in the proceedings of ECSCRM 2020-2021 (with corresponding DOIs of 10.4028/p-s1v84y and 10.4028/p-fnekfr). Finally, my research team has contributed to six separate inventions (of novel power semiconductor devices) by producing all simulation results on the predicted performance of these devices and preparing dedicated internal invention disclosures for each device. Our industrial partners now need to submit patent applications for each of these inventions. |
| Collaborator Contribution | Our collaborators have contributed to this project mostly by providing experimental data, which our group could then use in order to improve the accuracy of our computer models of the investigated designs. In addition, they have made various design suggestions and have also participated in the analysis of the simulation results produced by my research group during our technical meetings. |
| Impact | Our collaboration is centred primarily on the design and evaluation of the performance of Silicon Carbide power semiconductor devices. Thus, our partnership lies solely in the field of Electrical Engineering (with subfields of Semiconductor Engineering and Power Electronics) and is hence not multi-disciplinary. Otherwise, the specific outputs of our collaboration are: 1. 3 Four conference papers have been presented and accepted for publication in the proceedings of three separate conferences (one at ISPSD 2021 (with a DOI of 10.23919/ispsd50666.2021.9452208), two at ECSCRM 2020-2021 (with DOIs of 10.4028/p-s1v84y and 10.4028/p-fnekfr), and a final one at ISPSD 2022 (with a DOI of 10.1109/ISPSD49238.2022.9813631). 2. 6 (Six) invention disclosures have been prepared; patent applications still need to be submitted by our collaborators. 3. Numerous simulation results and corresponding analysis presented internally to our industrial partners during our technical meetings (confidential and hence not publicly available). |
| Start Year | 2019 |
| Description | SiC UNBMOS power semiconductor devices |
| Organisation | Toyota Motor Corporation |
| Country | Japan |
| Sector | Private |
| PI Contribution | My research group has contributed to this partnership primarily through studying and analysing the performance of a specific type (lateral and vertical Silicon Carbide UNBMOSFETs) of power semiconductor devices by means of computer simulations. We have produced computer models of the devices manufactured by our collaborators and have calibrated them to experimental data. The other members of my research group and I have subsequently used these models to predict the performance of other device designs. Based on these results, my research team has so far been able to publish two conference papers (DOI: 10.23919/ispsd50666.2021.9452282) in the proceedings of ISPSD 2021 and ISPSD 2022 (DOI: 10.1109/ISPSD49238.2022.9813617). The latter of these won the 2022 Ohmi award for the best paper presented at the conference. Another journal paper, which aims to further develop the ideas proposed in the aforementioned papers, is also presently being written and will soon be submitted for publication to Journal of Physics D. In addition, my research group has provided consistent advice on the design of new devices and on the methods for their investigation to our collaborators. |
| Collaborator Contribution | Our collaborators' key role in this project has been to provide access to experimental data and confidential information about their devices to us. In addition, they have also participated in the discussion of our simulation results during our technical meetings. |
| Impact | This collaboration focuses on the investigation of a specific type of power devices (in the field of Electrical Engineering) and is not multi-disciplinary. The key outcomes from it are as follows: 1. 1 Two published conference papers (DOI: 10.23919/ispsd50666.2021.9452282 and 10.1109/ISPSD49238.2022.9813617). The latter of these won the Ohmi award for the best paper presented at ISPSD 2022. 2. Numerous simulation results and design suggestions presented internally during the technical meetings with our collaborators (confidential information). |
| Start Year | 2019 |